It does have to do with cultural comfort, and the fact humans make hiring decisions instead of machines. In any case where human judgement has been replaced by measurement or judgement blind to demographics diversity has increased.

I know it's annoying when reality is non-optimized, but it doesn't take twisting ourselves into knots to explain why: humans make predictably bad decisions.

Citation needed for this one. It's pretty easy to to measure resume response rates, but I'd love to know how you would even go about measuring this one.

In any case where human judgement has been replaced by measurement or judgement blind to demographics diversity has increased.

This is simply false. For one example, take university admissions - after Grutter, universities have switched from objective systems to systems with human biases in order to increase demographic diversity.

Similarly, take a look at Ricci - the objective measurement system (firefighting tests) was replaced with a human judgement system in order to increase the number of minorities.

It's certainly true that diversity has increased in some cases, however - I believe women in classical music is the standard example.

There is a third, and much more likely scenario:

3) The employers' perceptions of Arabs is subtly colored by popular opinion and stereotypes as exists in pop culture and media. Much of these depictions and stereotypes appear harmless, but cumulatively add up to something substantial. Many of these depictions in media and culture are meant to be harmless, but subtly color your judgment anyways.

The employer is not out to discriminate against anyone, but their perception is colored, and therefore they make subconsciously sub-optimal choices sorting the stack of resumes.

That is a far more likely scenario, it just isn't as black and white, and isn't as interesting due to a noticeable lack of Bad Guys(tm).

I'll also disagree on the "betting against them is a winning strategy" thing. Your theory seems to be that bigotry will collapse in upon itself because they're denying themselves the objectively optimal choices. This doesn't stand up to actual example in history - Blacks were systematically oppressed until direct action was taken. The racists and bigots never caved in on themselves by cutting themselves out of a skilled labor pool, they succeeded until directly toppled.

What I mean is that as a founder of a business there is some people out there whoes skills are incorrectly priced by the market.

Yes, that was satire. Another possibility - improbable, I know - is that there are actually statistical differences between Frenchmen and Arabs. As a smart guy, you might enjoy this essay on pattern recognition:

http://isteve.blogspot.com/2010/10/war-on-pattern-recognitio...

Someone else posted George Akerlof's paper on asymmetric information:

http://web.usal.es/~emmam/Docencia/Modelizacion/papers/The%2...

Your second line of thought seems to approach "I'm not racist, arabs just actually are unenlightened bastards".

It should probably be noted when talking about the history of science during the so called dark ages that the Islamic Empire covered a very vast area. The society was homogeneous only in religion, not in 'race'.

My (limited) understanding is that their scientific collapse was the result of a series of (religious in nature) bans on various scientific activities (translation, dissection, etc) This seems to have had fair reaching effects on the viability of their Empire and has likely been the cause of the Muslim worlds comparatively limited scientific activity since then.

Therefore I would expect that the decline of Islamic science is not a symptom of a perceived performance deficiency among people with Arab names, but rather the cause. Considering the apparently solely cultural nature of the issue, any observed phenomenon could be subject to extremely rapid change and is likely to have a very significant number of outliers.